CN105612522A

CN105612522A - System and method for content-based medical macro sorting and search system

Info

Publication number: CN105612522A
Application number: CN201480053946.2A
Authority: CN
Inventors: Y·徐; 钱悦晨
Original assignee: Koninklijke Philips Electronics NV
Current assignee: Koninklijke Philips NV
Priority date: 2013-09-30
Filing date: 2014-09-22
Publication date: 2016-05-25
Also published as: JP6549563B2; CN113901079A; WO2015044852A1; US20160232329A1; US10198557B2; JP2016536659A; EP3053068A1

Abstract

When automatically populating medical report templates, insertable macros are indexed and searched not only my name or title but also by contents, such as keywords times, pre-defined terms for key information included in the macro, free texts in the macros, etc. When a unique macro is found, the system inserts the text of the macro into the report being generated. If multiple related macros are found, the system highlights the macros for user review. After the insertion of the macro into the template, the system identifies predefined terms and fills in the key information value(s). The system thus facilitates, e.g., radiologists' observation reporting procedure through an intelligent matching algorithm that facilitates finding a unique macro, which in turn aids in filling in report field instance values and optimizes radiology workflow.

Description

For the content-based grand classification of medical science and the system and method for search system

Technical field

The application is applied in medical report generation system especially. But, will understand, describedTechnology can also be applied to other report preparing systems, other medical scenes or other sorting techniques.

Background technology

Learn and read in workflow at conventional planning, after loading medical image, radiologist needsExamine image and understand it to generate the report for diagnosis or recommendation. Report is radiologistPatient's historical valuable source is observed and examined before reading image to the disease of record patient. MoreCome more ground, radiologist prefers to be absorbed on screen and uses exempting from such as speech microphoneThe instrument of carrying is given an oral account the key discovery during the report stage, and this also solves occupational stress and tired problem.And grand is conventionally to make for report and volume in the application of speed-up computation machine with oral account or together with keying inThe key character of the process of collecting.

At present, be grandly only pure free text and can only search for by title. The critical segment of informationCan be used to select from the list of standard medical term; For example, prostate can utilize kidney, spleenReplace in dirty and uterus etc. User can select one to be described in given anatomical structure place from optionThe discovery of observing. Like this, user can reuse grand by the various combination of value. But, routineMethod efficiently convenient by the key entry to keyword or voice from individual available grand the finding of hundreds ofUnique grand or template, and expect time, in grand, insert and edit some key message territories (word,Phrase). Unique grand in order to find, typical method is the free text using in grand title or titleSearch for. Conventionally it is grand that, radiologist only uses single narration. In the time that search is grand, conventional algorithm is adoptedBy the text matches that only uses macro name, therefore return potentially multiple grand; Many grand and current checksIrrelevant.

The application relate to the method that conveniently provides content-based with index and search for filling medical recordGrand new of the medical science of template and through improved system and method, it overcomes problem mentioned aboveAnd other problems.

Summary of the invention

According to an aspect, a kind of convenience to the medical science for filling medical report template grand based on interiorThe classification of holding and the system of search comprise: user interface, and by described user interface, user's typing oneIndividual or multiple words; And pretreatment module, the word of its analysis user typing and generation comprise according to instituteState the inquiry of the keyword that the word of user's typing generates. Described system also comprises matching module, described inMatching module: by described inquiry and multiple pretreated grand comparing; Generate the grand list of candidate, instituteState candidate and there is the grand of at least one word common with described inquiry grand comprising; And identification is for insertingUnique grand in described medical science template. Grand with respect to other candidates, described unique grand minimum number that comprisesObject word and with the word of the common maximum number of described inquiry.

According on the other hand, a kind of to filling, the medical science of medical report template is grand classifies and searches forMethod comprises: receive user input text; Analyze described user input text; Generation comprises according to instituteState the keyword that user input text generates; And by described inquiry and multiple pretreated grand enterRow relatively. Described method also comprises: generate and comprise having at least one word common with described inquiryThe grand list of grand candidate, and identification is for being inserted into the unique grand of described medical science template. With respect toOther candidates are grand, the described unique grand word of minimal amount and maximum number common with described inquiry of comprisingObject word.

According on the other hand, a kind of next unique grand from one group of grand identification with the shortest word matching algorithm of orderMethod comprise: the word that receives user input; And the word of the input receiving by intercepting is located in advanceThe word of user's input that reason receives, until each in the word of the input receiving is by its stem tableShow; Remove redundancy word; And remove stop-word. Described method also comprises: in described one group of grand middle knowledgeDo not there is the multiple grand of the common word of the word of at least one and described pretreated user input; AndIdentification is for being inserted into the unique grand of medical science template, wherein, and with respect to described multiple other in grandGrand, described unique grand word of minimal amount and common with the word of described pretreated user's input that comprisesThe word of same maximum number.

An advantage is, by content instead of only title or title to carry out index and search grand.

Another advantage is, reduces the medical report rise time.

Those skilled in the art will understand more of the present invention after describing in detail below reading and understandingFurther advantage.

Brief description of the drawings

The present invention can take the layout of various parts and each parts and the peace of various step and each stepRow's form. Accompanying drawing only should not be interpreted as limitation of the present invention for the object of diagram each side.

Fig. 1 illustrates conveniently carry out the grand content of index and automatically know by inputting one or more keywordsNot unique grand and automatically insert the framework of the predefined value in the filled in territory in medical report template.

Fig. 2 illustrates the method for carrying out the unique grand the shortest word matching algorithm of order of identification.

Fig. 3 illustrates according to the convenience of one or more aspects described herein and is filling in for usingGrand content-based classification in medical report etc. and the system of search.

Fig. 4 illustrate according to one or more aspects described herein to for filling medical reportThe classification that the medical science of template is grand and the method for search.

Detailed description of the invention

The present invention can be also based on comprising the grand of all key messages except macro name by providingContent, for the predefined term of key message and grand in the grand system of free text etc. search andMethod overcomes foregoing problems. Unique when grand when finding, system is inserted into grand text just and is generatedReport in. If find multiple relevant grand, outstanding described grandly examine for user of system. ToGrand be inserted in report template after, the predefined term of system identification and (one or more) are closedKey information value be filled in grand in. Therefore, described system facilitates the doctor of dept. of radiology by Intelligent Matching algorithmTeacher's visual report flow process is to find unique coupling grand, and its auxiliary inserting is reported territory example value and optimizationRadiology workflow.

As used herein " grand " represents to be inserted into the predefined text fragments in report. GrandIn textual description medical terminology, discovery/observation, observe character and diagnosis, normally complete sentence.The grand visual report flow process that can facilitate radiologist, reduce ambiguity, guarantee integrality and uniformity,Disclose best practices and optimize radiology workflow.

Fig. 1 illustrates convenient for example, by input (, key in or speak etc.) one or more passesKeyword (for example,, for the key message in grand predefined text) carrys out the grand content of index and automaticIdentify unique grand and predefined value is automatically inserted and in the filled in territory in report template, improved dept. of radiologyThe framework 10 of doctor's workflow. Framework 10 comprises pretreatment module 12 and content-based coupling mouldPiece 14. During pre-processing, grand 16 pretreated with by by voice label 18, territory example value 20With free text append to its revise grand, and modified grand stem module 24, the word of being imported intoKeyword is truncated to its corresponding stem by dry module 24, from grand redundancy word and the stop-word of removing. Also defeatedEnter in stem module being the word 26 giving an oral account and/or key in, it also extracts word by stem moduleDry. Stem module 24 also from the word 26 giving an oral account and/or key in remove redundancy word and stop words. SoAfter, that exports from stem module is pretreated content-based grand in from pretreated inputOne or more inquiries that word generates are stored in database 28. Pretreatment module 12 also comprisesMacroblock 30 based on contextual suggestion, it is advised based on contextual grand.

During content-based matching stage, comparator module 32 receives through pre-from database 28Reason grand and from the macroblock 30 of suggestion receive based on contextual suggestion grand and definite be through in advanceThe grand word that jointly still mates of the grand and suggestion of processing. If not, do not change at 34 places soBecome. Comparator module 32 can be used the content-based of for example the shortest word matching algorithm of order, three partsMatching algorithm (for example, using voice label, territory example value and free text) or any otherApplicable matching algorithm. If make definitely by comparator module 32, carry based on contextualWhat go out grand and that retrieve from database 28 content-based grandly has at least one common word, and asFruit is found and has the multiple grand of common word at 36 places, at 38 places, grand list is presented to user so.If definite find unique grandly at 40 places based on what made by comparator module 32, do at 42 placesGo out about any territory example value and whether need to insert grand determining. If not, so at 44 placesIn the report template that grand insertion is inserted. If territory example value need to be inserted as determined at 42 placesGrand, upgrade at 46 places the grand and thresholding of inserting in the report template of inserting so. As described aboveThe method of index relate to generate be used to search for grand in each grand content keyword. The method canTo adopt natural language processing (NLP) technology or to generate the technology such as semantic word as described below.

How following example has for example illustrated that index and storage are grand during pretreatment stage. Grand fixed belowJustice adopts the markup language that comprises three parts: voice label, key message (for example, territory example value)With free text. Voice label is to retrieve the bookmark name of given grand used general oral account (at thisIn example, " cystic mass "). Free text is the static part of the content in grand. For example:

" anatomical structure " is grand key message herein. In this example, use existing ontology tool orManually predefined term in anatomical structure, in this example, " anatomical structure " can utilize four choosingsItem is replaced: kidney, prostate, spleen and uterus.

Before in three parts are saved in to database, carry out pre-treatment step, during it, applyStem and stop words remove logic. Stem extraction module 24 is identified the root form of word, for example can be " to increase(enhanc) by force " stem extract " (enhancing) of enhancing " or " through strengthen (enhanced) ",Deng. To point out, in described method, root needn't be served as complete notional word and be existed. Stop-word bagDraw together some articles, conjunction etc., such as " there, be, should, one, have and, ", butBe not limited to this. After pretreatment, the description of above grand " cystic mass " becomes that " cystic mass is multipleAssorted kidney prostate spleen uterine abscess ", its expression is used to the grand index of content-based search. But before the content-based searching algorithm of application, it is at the beginning of the grand finite subset of minimizingBefore the step of grand list of beginning, wherein, use patient's DICOM (medical digital image with communicate by letter)Context in filter list.

Continue with reference to figure 1, Fig. 2 illustrates for carrying out the shortest word matching algorithm of the unique grand order of identificationMethod. But matching algorithm is used to identification to be had the lexical word of minimal number has maximum numberMate the unique grand or relevant grand of word. Unique grand be have comprise mate word multiple grand centreThe lexical word of peanut grand. The number of the word in grand is larger, and the chance of grand theme difference is larger,Even if it comprises common keyword. Therefore, the algorithm search grand candidate that is correlated with, and utilize couplingThe word of number of increase upgrade candidate pool, until it has identified coupling all search words and havingThe word of minimal amount grand, otherwise algorithm returns to coupling word relevant grand with maximum number.

According to Fig. 2, the word 26 of user's input of giving an oral account or keying in is extracted and pretreatment by stemTo remove redundancy word, stop-word etc., and the content-based pretreated grand database that is stored inIn 28, as what set forth in the pre-treatment step of Fig. 1. At 60 places, make about content-based grandIn N word whether mate determining of word in the word of pretreated user input, wherein, at the beginning of NBeginningization is to 1. If not, it is follow-up content-based to retrieve that method turns back to database 28Grand. If the word that (one or more) content-based grand and user inputs has common word, soGenerate grand candidate list at 62 places. At 64 places, make about any grand in candidate list and be whetherCommon the determining of two words of word with user's input. If not, at 66 places, candidate list is returnedReturn and present to user. If it is common that the grand word of inputting with user of one or more candidates has two words,N increment (for example to 3 etc.) and method turn back to 62 so, wherein, upgrade candidate list.Whether side by side or periodically, make all words of given candidate in grand at 68 places inputs userWord common determine. If not, method turns back to 62 so. If exist70 places, return to grand the shortest candidate with all the elements word (being unique grand) present to user.

Identification unique grand after, algorithm use be attached to the knowledge in unique grand different territory andAutomatically differential threshold and its value while mentioning as voice-based input. For example, if saying, user " separatesCut open structure spleen ", differential threshold title and territory example value and utilization value " spleen " filling so " dissectedStructure " territory. In another example, briefly " spleen " of user, and algorithm identified hasThis value as the territory (anatomical structure in this case: spleen) of one of preset list of its value andIts value is updated to " spleen ". In another example, if user says " measurement result 1.6 ", profitUpgrade " measurement result " territory by user's value of providing, described user's value of providing is word " measurement result "Numeral (" 1.6 " in this case) afterwards.

According to another example, if order " grand cystic mass kidney " as input word be uttered orKey in, system is initiated for the grand search of unique coupling, and generates the word comprising for this orderThe query vector of " cystic mass kidney ". Then, grand by query vector and database of systemMultiple vectors are compared. The minimum word that will have in its vector still has with query vector commonGrand (grand with respect to other) in the database of the word of maximum number is identified as unique grand. At this time,To insert in report template for grand predefined text, and the utilization of " anatomical structure " key messageFor example, " kidney " in grand replaced. Will appreciate that, described search and Index Algorithm can also be usedMake template searching algorithm.

Grand is below the predefined text of describing for grand " cystic mass prostate ", and it will be inserted inUnder " discovery " section.

The complicated cystic mass of measuring [] taking mm as unit is present in prostate.

Aforementioned is that " semanteme " is grand, and it is the structure of the content that comprises predefined type of some terms in termChange grandly, as shown in following example, wherein, square brackets represent different predefined terms.

The complicated cystic mass of measuring [measurement result] taking mm as unit is present in [anatomical structure].

The part of grand " anatomical structure " has the territory example value of describing the territory in grand above. For this territory, useFamily can the selectable some options of tool. For example, " anatomical structure " can have such as kidney, prostatitisSome value options in gland, spleen or uterus. User selects to be described in given anatomical position place and observesAn option of discovery. By this way, user can pass through the different combination weight of territory example valueUse again grand.

How index structureization is grand for following example illustration. Following macrodefinition adopts the mark that comprises three partsNote language: voice label, territory example and free text. Voice label be retrieval this grand " title " (In this example, it is " cystic mass "). Free text is the static part of the content in grand.

< grand key=" cystic mass " title=" cystic mass "

Voice label=" cystic mass " anatomical structure=" territory example: anatomical structure "

Occurrence type=" lump " found code=" http://www.radlex.org/RID/RID3874 " >

<free Wen Ben>Complicated cystic mass is measured</free Wen Ben>

<free Wen Ben>Be present in</free Wen Ben>In

< territory example key=" organ expands. anatomical structure " and title=" anatomical structure " voice label=" dissectStructure "/>

<free Wen Ben>Suspicious for abscess.</free Wen Ben>

</Hong>

The position that keeps semantic thresholding with the instruction of co-domain example. For example, territory example " anatomical structure " refers to territoryValue collection " organ expands. anatomical structure ", it can be defined as follows:

</Yu>

Therefore, system and method described herein conveniently carry out pretreatment, based on contextual pre-mistakeFilter is to advise grand and mate with identification and retrieve maximally related grand with the content-based of patient data. In advanceProcessing module 12 is by the database 28 in the grand structure that is stored in searchable and editor. Based on upperFiltration is below used the DICOM information of present image research that grand list is narrowed. Then, carry outContent-based coupling is suitable grand to find.

Fig. 3 illustrates according to the convenience of one or more aspects described herein and is filling in for usingGrand content-based classification in medical report etc. and the system 100 of search. System comprises and can operateBe coupled to the processor 102 of memory 104 and user interface 106. Processor 102 moves computerExecutable instruction and memory 104 store computer executable instructions (for example, routine, program,Algorithm, software code etc.) for carrying out various function described herein, method, flow process etc.Deng. In addition, as used herein " module " represent set of computer-executable instructions, software code,Program, routine or can actuating unit for other computers of carrying out described function, as abilityThe technical staff in territory by understand.

Memory can be that the computer-readable of the control program of storage such as disk, hard-drive etc. is situated betweenMatter. The common form of non-transient state computer-readable media comprises for example floppy disk, flexible disk, hard disk, magneticBand or any other magnetic-based storage media, CD-ROM, DVD or any other optical medium,RAM, ROM, PROM, EPROM, FLASH-EPROM, its modification, other storage coresAny other tangible medium that sheet or chuck or processor can read and move from it. On thisHereinafter, system described herein can realize one or more all-purpose computers, (one or manyIndividual) special-purpose computer, programming microprocessor or microcontroller and peripheral integrated circuit component, ASIC orOther integrated circuits of person, digital signal processor, hardwire electronic equipment or logic circuit for example divideVertical element circuitry, programmable logic device for example PLD, PLA, FPGA, graphics card CPU (GPU)Or PAL etc.

Memory stores is multiple grand 16 (referring to for example based on its content indexing by pretreatment module 12Fig. 1). Pretreatment module 12 is analyzed and is comprised user's input of the oral account receiving and/or connect via userMouth 106 key entries that receive 26 and generated query are grand to search. Processor Operating match module 14And memory stores matching module 14 (referring to for example Fig. 1) is with based on being given birth to by pretreatment module 12The query search becoming is grand. Matching module 14 generates the grand list 62 from its unique grand candidate of identification, as closedDescribed in Fig. 1 and 2. Also being stored in what in memory and by processor, move is territory example valueModule 108, grand territory and the insertion literary composition of coupling that its part of automatically identifying input text 26 belongs toThis and the territory of identifying. Once it is unique grand to have utilized identified territory example value to fill, grand insertionModule 110 is by filled grand being inserted in report template 112, and it is then via user interface 106Present to user. In an example, user interface 106 presents the relevant district that is inserted into template by grandPredefined text in section and outstanding revisable territory in grand are examined for user.

Memory is storaged voice identification module 114 also, and sound identification module 114 is moved to hold by processorThe pass through speech recognition and the speech data that are received from user by user interface 106. According to retouching hereinOne or more features of stating, speech recognition is used to navigation by recognition voice command. For example, system existsIn " wait " state, start and switch mode when receive order by user interface. For example, asFruit order starts with " grand ", system in " grand " state, and etc. grand keyword to be searched.In order to continue this example, in voice command " grand cystic mass prostate ", keyword is " capsuleLump prostate ". If system can not find relevant grand, it remains in " grand " state, waits forNew keywords. Identity logic can be suitable for by its classification model and with about grand described herein" template " order that mode is searched for.

Can also provide order to insert territory example value. For example,, after selection is grand, if user says" anatomical structure liver ", system utilization value " liver " filling " anatomical structure " territory. Keyboard inputSystem is moved in free text editor state. That kind, any keyed in input is insertedIn report region. System can also be by saying that " adding bookmark 5 " is by bookmark (to the observation of image)Be linked to grand, its create in image bookmark #5 in current check and report current selected grand itBetween permalink. User can upgrade and existingly grandly (announces the first territory by changing thresholding in grandName, is followed by its territory example value). Also provide navigation speech order for grand management. For example, orderAlso be suitable for filtering by anatomical structure grand, the different section in all grand and navigation reports is shown.

Fig. 4 illustrate according to one or more aspects described herein to for filling medical reportThe classification that the medical science of template is grand and the method for search. At 140 places, receive user input text. CanKey in text, (for example oral account) speech data, etc. At 142 places, analysis user input text withIdentification is included in the stem of the word in input text etc. At 144 places, generate and be included in user's inputThe inquiry of the keyword stem of identifying in text. At 146 places, will inquire about with multiple pretreatedGrand comparing. Pretreatedly grandly extracted keyword to be reduced to its corresponding stem by stemAnd redundancy word and stop words are removed. At 148 places, generate comprise have with inquiry at least commonThe grand list of grand candidate of word or stem. At 150 places, unique grand being identified for being inserted intoIn medical science template. Described unique grand word comprising with respect to the grand minimal amount of other candidates, and relativelyThe word of and maximum number that described inquiry common grand in other candidates.

In one embodiment, for example pass through one of voice label, territory example value and free textOr multiple appending to as multiple grand according to grand content indexing about the described macroradical of aforementioned figures. CanOne or more revisable in grand for the insertion identification coupling of at least a portion of input textTerritory, and input text can be inserted in the territory that (one or more) identify, fill grand,Then it present to user via user interface. Extraly or alternatively, grand predefined text canTo be inserted in the relevant section of medical science template, and one or more revisable territory in grand is passableBe highlighted for user and examine. In another embodiment, speech recognition is used to analyze and connects via userThe speech data that mouth receives from user and the input that is provided for identifying pretreated speech dataWord.

With reference to some embodiment, the present invention is described. Other people are in reading and understand aforementioned detailed theoryAfter bright, can expect amendment and modification. The present invention be intended to be understood to include all such amendments andModification, as long as within it falls into the scope of claim or its equivalence.

Claims

A convenience grand content-based of the medical science for filling medical report template classified andThe system (10,100) of search, comprising:

User interface (106), by described user interface, the one or more words of user's typing;

Pretreatment module (12), the word of its analysis user typing and generation comprise according to described user's recordThe inquiry of the keyword that the word entering generates;

Matching module (14), described inquiry and multiple pretreated grand comparing are generated bag by itDraw together the grand list of grand candidate (62) with at least one word common with described inquiry, and identification is usedUnique grand (40) in being inserted into medical science template (112);

Wherein, grand with respect to other candidates, described unique grand (40) comprise minimal amount word andWord with the common maximum number of described inquiry.
2. system according to claim 1, also comprises the database of storing multiple grand (16),Described multiple grand by described pretreatment module (12) based on described multiple grand contents and indexed.
3. system according to claim 2, wherein, described pretreatment module is passed through voice markSign, one or more in territory example value and free text append to described grand come grand described in index(16)。
4. according to the system described in any one in aforementioned claim, also comprise territory example value module(108), described territory example value module is identified belonging at least partly of input text (26) automaticallyMate grand territory, and the described part of insertion input text is described grand to fill with the territory of identifying.
5. system according to claim 4, also comprises grand insert module (110), described grand insertingEnter module by the filled grand report of presenting to described user via described user interface (106) that is inserted intoSlide former (112).
6. system according to claim 5, wherein, described user interface (106) by insertThe grand predefined text entering is presented in the relevant section of described medical science template, and outstanding described grandIn one or more revisable territory examine for described user.
7. according to the system described in any one in aforementioned claim, also comprise sound identification module(114) language that, described sound identification module analysis receives from described user via described user interfaceSound data and input word is offered to described pretreatment module.
8. to the grand method of classifying and searching for of the medical science for filling medical report template, wrapDraw together:

Receive user input text;

Analyze described user input text;

Generation comprises the inquiry of the keyword generating according to described user input text;

By described inquiry and multiple pretreated grand comparing;

Generation comprises the grand list of grand candidate (62) with at least one word common with described inquiry;And

Identification is used for being inserted into unique grand (40) of described medical science template (112);

Wherein, grand with respect to other candidates, described unique grand (40) comprise minimal amount word andWord with the common maximum number of described inquiry.
9. method according to claim 8, also comprises that to carry out index according to described grand content multipleGrand (16).
10. method according to claim 9, wherein, described in index grand (16) also comprise byIt is described grand that one or more in voice label, territory example value and free text append to.
Method described in any one in 11. according to Claim 8-10, also comprises described in automatic identificationThe grand territory of the coupling belonging at least partly of input text (26), and by described in input textDescribed grand to fill in the territory that partial insertion is identified.
12. methods according to claim 11, also comprise filled grand being inserted into via useFamily interface (106) is presented to user's medical report template (112).
13. methods according to claim 12, also comprise inserted grand predefined textBe inserted in the relevant section of described medical science template, and outstanding described one or more repairing in grandThe territory changing is examined for described user.
Method described in any one in 14. according to Claim 8-13, also comprises and analyzing via userThe speech data that interface receives from user and the input word of identifying described speech data is providedGive pretreatment module to carry out pretreatment.
15. 1 kinds carry processor (102) or the computer-readable medium (104) of computer program,The one or more processors of described computer program control are to carry out as any one in claim 8-14Described method.
16. 1 kinds for coming from one group of unique grand side of grand identification with the shortest word matching algorithm of orderMethod, comprising:

Receive the word (26) of user's input;

Carry out in the following manner the word of user's input that pretreatment receives:

The word of the input that intercepting receives, until each in the word of the input receiving is by its wordDry expression;

Remove redundancy word;

And remove stop-word;

The word in described one group of grand middle identification with at least one and pretreated user input is commonWord multiple grand; And

Identification is used for being inserted into unique grand (40) of medical science template (112), wherein, and with respect to instituteState multiple other in grand grand, described unique grand (40) comprise and described pretreated user's inputThe common word of minimal amount and the word of maximum number of word.
17. methods according to claim 16, also comprise according to described grand content and come index instituteState one group grand.
18. according to the method described in any one in claim 16-17, wherein, described in index grand alsoComprise one or more in voice label, territory example value and free text are appended to described grand.
19. according to the method described in any one in claim 16-18, also comprises described in automatic identificationThe grand territory of the coupling belonging at least partly of input text (26), and by described in input textDescribed grand to fill in the territory that partial insertion is identified.
20. methods according to claim 19, also comprise filled grand being inserted into via useFamily interface (106) is presented to described user's medical report template (112).